This file provides replication information for 

HABIT PERSISTENCE, NONSEPARABILITY BETWEEN CONSUMPTION AND LEISURE,
OR RULE-OF-THUMB CONSUMERS: WHICH ACCOUNTS FOR THE PREDICTABILITY
OF CONSUMPTION GROWTH?
Michael T. Kiley*

The Review of Economics and Statistics, August 2010, 92(3): 679-683

0. All calculations performed with TSP.  For specific version, see
*.OUT files (which report full output).

1. DATA

The data are provided in two portable tsp files.

RAWDATA.TSP provides the raw data (with some documentations) for the
data used subsequently.  It can be used to generate RAWDATA.TLB -- the
input for clndata2.tsp (which transforms the data and creates CLNDATA2.TLB).

CLNDATA2.TLB is the main input data file for subsequent programs,  It
can also be reproduced using the portable TSP file CLNPORT.TSP

Note: For those unfamiliar with TSP, the portable TSP files above
include reports of the data in ASCII format.

2. Programs and output reported in the paper.

(NOTE: Output files below include full replication code)

Table 1 replication

The estimates for table 1 are using the code reported in weak_inst_sets3.out
Column 1 is estimated on line 38.  Column 2 is estimated on line 37.

Table 2 replication

The estimates for table 2 are using the code reported in
Column 1: weak_inst_sets1.out (line 44).
Column 2: weak_inst_sets2.out (line 44, refer to eiggcd for the appropriate eigenvalues).

Figure 1 replication

The Anderson Rubin (AR) statistics for instrument set 1 are computed
using the code reported in weak_inst_sets1.out, lines 60-72. Line 88
reports the output for those values of the AR statistic that do not
reject the null hypothesis.

The Anderson Rubin (AR) statistics for instrument set 2 are computed
using the code reported in weak_inst_sets2.out, lines 50-60. Line 74
reports the output for those values of the AR statistic that do not
reject the null hypothesis.

Table 3 replication

The estimates in table 3 are using the code in weak_inst_sets5.out
Column 1 is estimated on line 42. Column 2 is estimated on line 41.

Figure 2 replication

The Anderson Rubin (AR) statistics for instrument set 1 are computed
using the code reported in weak_inst_sets5.out, lines 45-60. Line 75
reports the output for those values of the AR statistic that do not
reject the null hypothesis.

The Anderson Rubin (AR) statistics for instrument set 2 are computed
using the code reported in weak_inst_sets4.out, lines 45-60. Line 75
reports the output for those values of the AR statistic that do not
reject the null hypothesis.
